Skip to content
View m-ulmestrand's full-sized avatar
:electron:
Doing AI stuff
:electron:
Doing AI stuff
6 followers · 1 following
  • Gothenburg, Sweden

Achievements

Achievement: QuickdrawAchievement: Pull Sharkx2

Achievements

Achievement: QuickdrawAchievement: Pull Sharkx2

Block or report m-ulmestrand

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
m-ulmestrand/README.md

  • 👋 Hi, I’m @m-ulmestrand, but you can call me Mattias.

  • 👀 I do a lot of Machine Learning in my spare time as well as at work. I'm especially interested in Graph Neural Networks. I find them useful and fascinating, since a lot of physical structures can be modelled as graphs. I have a repository open where I have used a GCN to learn collision physics in multi-particle collisions. I also made a self-learning game with LSTM and Deep Q-learning, where agents learn to avoid the predator (player) in real time. Another project I did features self-driving cars with Deep Q-learning and LiDAR-like detectors

  • 🌱 I work as a Machine Learning and Computer Vision engineer at Taigatech. Previously I studied the Master's programme Complex Adaptive Systems at Chalmers University of Technology where I predominantly took courses in Machine Learning and Computational engineering. Before that I studied Engineering Physics for three years. I also conducted exchange studies to ETH Zürich, where I continued with the same profile. Having a background in both Computational engineering and ML, I like to do projects where I get to combine the modelling-intensive character of Computational engineering and the data-centred approach of ML. Most of my projects have such an approach, where I model an environment from the ground up and let an ML algorithm learn from it.

  • 💞️ I’d love to collaborate on ML projects, especially with applications for physics, programming or games.

  • 📫 Reach me at m.ulmestrand@gmail.com.

Pinned Loading

  1. self-driving-ai self-driving-ai Public

    This project explores making a model of a car self-driving. The model is controlled by a neural network which is trained with Q-learning. The entire environment, from collision detection to buildin…

    Python 5 1

  2. gcn-collisions gcn-collisions Public

    This Graph Convolutional Network learns to simulate collision responses for particles. Speedups can be seen for large numbers of particles. The network can also learn to account for symmetry-breaki…

    Python 3

  3. interactive-AI-game interactive-AI-game Public

    This game lets the player control a predator, whose objective is to devour prey agents. The agents learn to avoid the player from experience in real time. The agents are controlled by an LSTM which…

    Python 7 2

  4. art-ificial-intelligence art-ificial-intelligence Public

    This is a master's project in advanced machine learning with neural networks. We investigated CycleGANs for generating art of different styles, compared different architectures and improved upon th…

    1